No Priors: Artificial Intelligence | Technology | Startups NVIDIA’s Jensen Huang on Reasoning Models, Robotics, and Refuting the “AI Bubble” Narrative
1019 snips
Jan 8, 2026 
Jensen Huang, CEO of NVIDIA and a visionary in accelerated computing, engages in a dynamic discussion about AI and robotics. He reveals surprising advancements in reasoning and the profitability of inference tokens. Huang argues that AI will enhance job creation through new industries and solve labor shortages with robotics. He also addresses the importance of open source in maintaining competitive advantages and refutes the idea of a looming AI bubble. The conversation covers optimism in geopolitics and the future of diverse AI applications.
AI Snips
Chapters
Transcript
Episode notes
Watch For Regulatory Capture
- Beware regulatory capture where incumbents push rules to stifle competition and slow innovation.
- Scrutinize whose interests proposed regulations serve before supporting them.
Falling Token Costs Enable Automated Oversight
- Lower marginal costs of AI make widespread monitoring by other AIs feasible, improving overall safety and oversight.
- AI systems are likely to be surveilled and regulated by many automated agents, not act alone.
Compounding Gains In Compute And Algorithms
- Compute, algorithms, and model improvements compound annually, often yielding 5–10x gains per generation.
- Huang predicts token-generation cost could fall by orders of magnitude within a decade, enabling many new applications.